OpenAI Urges Mandatory National AI Safety Regulations in the US

OpenAI on Wednesday publicly called for mandatory, capability-based national artificial intelligence safety regulations in the United States, urging Congress to act before its recess. The appeal follows incidents in which some of OpenAI's own AI agents exhibited uncontrolled behavior during testing, raising concerns about safety risks in advanced agent systems.

Chris Lehane, OpenAI's head of global affairs, said in a blog post that the prospect of AI accelerating AI development demands more than voluntary commitments, and that the US needs binding national rules that can evolve with the technology. The company proposed requirements including testing standards for the most advanced AI systems, independent evaluations, cybersecurity protections, and incident reporting rules.

The US currently lacks a federal AI legal framework, while multiple states have introduced their own legislation. OpenAI said it will continue to support state-level AI bills until Congress acts. The company also backed four California AI regulation bills, two of which have been signed into law, establishing a legal framework for independent third-party assessments and audits of AI systems.

OpenAI added that fully autonomous recursive self-improvement has not yet been achieved and should not be pursued unless it can be done safely. The company also said that industry standards for AI regulation need to extend beyond US borders, and that compatible international standards will be essential.